Managing mine-contact water effectively and sustainably in (sub)arctic regions is crucial for expanding mining activities. The demand for cost-effective (semi-)passive water treatment that relies on natural chemical and biological processes and can withstand challenging weather conditions is increasing. This study investigated the ability of four pilot-scale bioreactors inoculated with locally sourced bacteria andaffected by a freeze-thaw cycle to effectively remove selenium and antimony. The bioreactors were operated at a Canadian subarctic mine for a year. Two duplicate bioreactors were installed in a heated shed maintained at 5°C over winter, while two other duplicates were installed outdoors and left to freeze. The removal rate of selenium and antimony was monitored weekly, while a genomic characterization of the microbial populations in the bioreactors was performed monthly. The bioreactors successfully removed selenium and antimony over the year, demonstrating their ability to manage freeze-thaw cycles. The overall percentage of selenium and antimony removal was similar in the outside and inside bioreactors, apart from the spring thawing period, when removal in the outdoors bioreactors was slightly lower. The dominant taxonomic groups of microbial populations were similar in all bioreactors, with slight variations observed in their relative abundance over time. The microbial population composition was consistent and re-established quickly after spring thaw in the outside bioreactors. This study demonstrated that the removal capacity of bioreactors inoculated with locally sourced bacteria was not largely affected by a freeze-thaw cycle, highlighting the strength of using local resources to design bioreactors in extreme climatic conditions.
Research Article
Semi-passive pilot scale bioreactors metal(loid) removal performance response to seasonal freeze-thaw cycle
https://doi.org/10.21203/rs.3.rs-3783461/v1
This work is licensed under a CC BY 4.0 License
Journal Publication
published 03 Sep, 2024
Read the published version in Mine Water and the Environment →
You are reading this latest preprint version
Sulfate reduction
mine-contact water
subarctic
genomics
semi-passive treatment
Use of metal(loid)s in clean and renewable energy technologies, daily used products and essential infrastructure, generates a vital need for mineral resources (Coulomb et al. 2015; McNulty and Jowitt 2021). The (sub)arctic region is rich in strategic mineral resources and is becoming more accessible thanks to infrastructure development, expanding significant interest from the mining sector (Haley et al. 2011; Tiainen et al. 2015). However, when not adequately managed, mining activities can present environmental concerns (Sengupta 2021). One major challenge is controlling the release of mine-contact water (MCW) into the environment. Unmanaged MCW containing metal(loid)s and other inorganic constituents at concentrations higher than the regulated limits might result in deleterious effects on downstream aquatic environments when MCW is released (Wright et al. 2015; Leppänen et al. 2017). Thus, the development of sustainable, efficient, and cost-effective water treatment systems is required for mineral exploitation to expand in (sub)arctic areas.
Active chemical water treatment systems are still preferred for operating mines (Qin et al., 2019). However, these systems rely heavily on infrastructure, labour, and chemical reagent utilization, limiting their sustainability in the long term (Ness et al. 2014). Thus, passive and semi-passive water treatment systems are considered viable alternatives in closure scenarios because they are more economically and environmentally friendly (Martin et al. 2009). Semi-passive water treatment mimics natural chemical and biological processes associated with some periodic active management, such as carbon source addition (Martin et al. 2009). They require little maintenance and energy resources once in operation (Younger et al. 2002; Johnson and Hallberg 2005). Among semi-passive water treatments, bioreactors (BRs) have been gaining interest in the mining sector (Neculita et al. 2007; Habe et al. 2020). BRs are constructed systems that support a biologically active environment able to sustain processes favouring the removal of specific inorganic constituents to reach levels that meet environmentally safe aquatic life standards.
Use of BRs has been proven efficient at the bench and pilot scale, in temperate climatic conditions and for temperatures as low as 5°C (Neculita et al. 2007; Nielsen et al. 2018b, a; Habe et al. 2020). However, their application in the (sub)arctic regions is challenged by the distinctive cold climatic conditions and limited resources available in the North. The climatic conditions in the (sub)arctic region are characterized by freeze-thaw cycles: long, cold and dry winters, with temperatures below freezing measured from late September to early June, followed by summer periods characterized by an increase in temperature and long daylight exposure. These patterns lead to multiple freezing and thawing events during the fall and the freshet. While during winter, ditches and watercourses that collect MCW often freeze. BRs are usually designed to be installed above ground and treat surface waters, which would result in freezing of the BRs and MCW over winter. In addition, low temperatures are known to limit bacterial activity (Price and Sowers 2004). One of the main challenges faced by the utilization of BRs in the (sub)arctic is, thus, the ability of the microbial population to survive freezing events and to become metabolically active again after the thaw. Disturbance to the bacterial population could decrease the capability of BRs to remove the targeted contaminants properly.
Previous work investigated the ability of BRs inoculated with locally sourced bacteria and installed in a heated shed (T°C ≥ 5°C) to remove contaminants (cadmium (Cd), zinc (Zn), sulfate (SO42−)) from MCW on a closed mine in the Yukon Territory in Canada (Nielsen et al. 2018a). The BRs removed Zn and Cd from MCW consistently over a 14-month period (Nielsen et al. 2018a). While this study showed that indigenous bacteria successfully supported processes for MCW decontamination, it did not investigate the effect of freeze-thaw cycles on the microbial population composition of the BRs and how this might affect removal efficiency. To the best of our knowledge, no study has investigated the effect of freeze-thaw cycles on the operation of BRs at the pilot scale and on their removal efficiency, a key question for the future of sustainable mining exploitation in the sub(arctic).
This study aimed to investigate the ability of semi-passive BRs to remove selenium (Se) and antinomy (Sb) from MCW under on-site conditions. Four BRs, inoculated with locally sourced bacterial populations, were installed on an operating mine site in the Yukon Territory, Canada (> 60°N). The first set of BRs, BR1 (BR1a and BR1b acting as duplicates), was set up in a shed where temperatures were maintained at ~ 5°C over winter to prevent the BRs from freezing. They operated at outside temperatures from spring to fall. The second set of BRs, BR2 (BR2a and BR2b), was set up outside the shed and left to freeze over winter and thaw in spring. It was possible to monitor the ability of these BRs inoculated with local native bacteria to remove Se and Sb while also comparing the performance of BRs that were inactive over winter (BR2) with that of BRs that did not freeze (BR1). In this study, the concentrations of organic and inorganic constituents in the MCW feed and BRs’ effluents were measured over 13 months, and the BRs’ bacterial populations were monitored once every month.
2.1 Experimental setup
Four 200 L pilot-scale BRs were installed at the remote operating Eagle Gold Mine in central Yukon Territory, Canada. Each BR vessel was an open-top 55-gallon polyethylene drum covered with a lid and a steel lever lock (Uline, Milton, Canada). Duplicate BRs (BR1a and BR1b) were placed within a shed that was heated to maintain a minimum temperature of 5°C over winter and left to fluctuate with outside temperature when over 5°C. Duplicates BR2a and BR2b were installed outside of the shed where they were exposed to freezing temperatures. MCW was fed into all BRs from the same feed tank that was situated inside the shed. MCW and the carbon source were fed into the BRs through inlet pipes installed near the bottom of each vessel. The MCW flowed from the bottom to the top of the BR vessels to ensure anoxic conditions. Oxidation-reduction potential (ORP) was measured approx. half way up and monthly in the BRs to check and confirm reducing conditions. The outflow from the top of each BR went to its own collector drum. Two collector drums inside of the shed remained open, while the other two collector drums outside of the shed were covered with lids and steel lever locks to prevent particles from entering the vessels. Temperatures inside the BRs were monitored using Oakton Temperature Probes 5+ installed at the top of the BRs, reaching the inside through a hole drilled in the BRs’ lid and sealed with silicone to prevent air ingress. Thirty-psi pressure relief Apollo valves were also installed on the lid to prevent pressure higher than 30 psi in the BRs (Figure 1).
Each BR was filled with 20% v/v locally sourced, shredded spruce wood chips 2.5-5 cm in size and 20% v/v inoculum. The inoculum was a mixture of sediment samples from two different wetlands in the same drainage of and close to Eagle Gold Mine, locations 0459110E 7100937N and 0458249E 7099695N, respectively. The wood chips and inoculum were mixed thoroughly in the BRs that were subsequently filled with 170L MCW. A passive sampling method was used to collect the microorganisms growing inside the BRs. The microorganisms sampled were assumed to be representative of the bacterial communities within the BRs, originating from either the inoculum or the MCW entering the BRs. Sampling bags were made with light cotton material and filled with the same sediment used to inoculate the BRs. Twelve sampling bags (one for each monthly sample) were suspended in the middle of each BR, with fishing lines fed through a hole in each BR lid at the beginning of the experiment. The hole was then sealed with silicon to prevent air ingress into the BRs.
The MCW feed for the BRs was restocked periodically from two different sources at Eagle Gold Mine, referred to as source 1 and source 2. From September 7, 2019 until November 11, 2019, and after May 12, 2020 to September 27, 2020, the MCW was collected from source 1, a pond above ditch A on the Platinum Gulch Waste Rock Storage Area. Between those dates, the MCW was sampled from source 2, a settling pond collecting MCW from ditch A.
2.2 Operation and experimental design
The BRs were operated with a two-week hydraulic retention time (HRT) based on a previous study with similar MCW parameters and carbon sources (Nielsen et al. 2018a). The MCW flow rate, controlled by peristaltic pumps (Cole Palmer Masterflex L/S Standard Digital Drives), was 10.6 mL/min.
The carbon source was molasses (Crosby’s 100% Natural Fancy Molasses) since this was used successfully in previous studies to promote the growth of sulfate reducing bacteria (SRB) (Muyzer and Stams 2008). In addition, molasses support diverse bacterial community (Zhao et al. 2010), thus, favoring a better resilience of the bacterial community in the BRs to environmental perturbations (Allison and Martiny 2008; Ayala‐Muñoz et al. 2021). The molasses solution was prepared biweekly. The amount of molasses added was based on the stoichiometric requirement (Eq. 1). Considering a maximum sulfate concentration in the MCW feed of 100 mg/L, the molasses concentration was fed into the BRs at 25 mg-C/L. Molasses solutions were pumped into the BRs for 22 minutes per day at a flow rate of 6 mL/min to achieve the targeted calculated total organic carbon (TOC) concentration per day.
Equation 1:
Given the site’s remoteness and the climatic conditions during winter, effluent samples were collected weekly from the effluent drums after mixing by stirring. The total volume of effluent collected in each drum over the previous week was recorded to verify accurate pumping rates and HRT. The conductivity and pH of the MCW and effluents were recorded weekly using an Oakton PCD650m pH meter equipped with a double junction Cole Palmer Ag/AgCl electrode. Samples for TOC were preserved with sulfuric acid (2%, v/v). Samples for heavy metals (HM) and metalloids were preserved with nitric acid (2%, v/v). Every month, one passive microbiological sample bag was removed from each BR. Sampling bags were frozen by liquid nitrogen in a thermo-flask (Thermo Fisher Scientific container 2122) immediately after collection and transported from the field to the laboratory for storage at -86°C until DNA extraction. Liquid samples were analyzed by ALS, Burnaby, for heavy metals and metalloids (EPA 200/6020A) and TOC (APHA 5310B). Sulfate analysis was performed following the protocol described by Nielsen and collaborators (Nielsen et al. 2018a).
2.3 DNA extraction, sequencing, and bioinformatics
Genomic DNA was extracted from 0.25 g subsamples taken from homogenized passive sample bag contents using the DNeasy PowerSoil Pro Kit (QIAGEN) following the manufacturer’s instructions. DNA concentrations were measured using an Invitrogen™ Qubit™ 3.0 Fluorometer (Thermo Fisher Scientific). The purity of the extracted DNA was measured using a NanoDrop ND-2000 Ultraviolet–visible spectrophotometer (NanoDrop Technologies) at 260 nm for nucleic acid. To ensure the purity of extracted DNA, absorbance was also measured at 230 nm for organic contamination and 280 nm for protein contamination. For pure DNA samples, absorbance ratios should be A260/A280 1.8-2.0 and A260/A230 2.0 (Armbrecht, 2013). If DNA did not pass QC, it was purified using the DNeasy PowerClean Pro Cleanup kit (QIAGEN).
The 16S rRNA variable region V4 to V5 was amplified using primers 515F-Y and 926R (Parada et al. 2016; Walters et al. 2016) and sequenced on an Illumina MiSeq in the Biofactorial Core Facility in the Life Sciences Institute of the University of British Columbia. The raw sequence data were converted into amplicon sequence variants (ASV) by applying the open-source software package DADA2 (Callahan et al. 2016) for modelling and correcting Illumina-sequenced amplicon errors, using the Qiime2 suite of tools (Bolyen et al. 2019). A frequency table with relative abundance information for each ASV was produced. The taxonomic classification of each ASV was assigned by aligning the representative sequence to the Silva SSU database (version 138.1) (Warnow 2015). All data were visualized in R or Excel. For a- and b-diversity measures, all samples were subsampled to the lowest coverage depth and standard indices were calculated in Qiime2. Raw sequencing data are available under NCBI BioProject ID PRJNA924841.
2.4 Carbon leaching experiments
To assess the TOC leaching potential of the wood chips used in the BRs, two duplicate 2L columns filled with 20% v/v wood chips (equivalent to 72g of wood chips per column) were installed in the laboratory at room temperature. DI water was pumped from the bottom to the top of the columns by a peristaltic pump (Cole Palmer Masterflex L\S Standard Digital Drives) for 35 days. The pump was set at 0.08 RPM to achieve an HRT of 14 days, similar to the one used in the pilot scale BRs. Weekly batches of leachate effluent were collected in 2 L jars. Homogenized samples were preserved with sulfuric acid (2% v/v) and sent to ALS, Burnaby, for TOC analysis.
3.1 Operating conditions
All BRs were operated from September 7, 2019 to September 27, 2020. From mid-October 2019 to mid-May 2020, referred to as the “winter” period, BR2a and BR2b were either fully or partially frozen and no MCW was flowing through them. It was expected that freezing and thawing occurred not at once but over a period of successive freeze-thaw events. The inside BRs continued to operate throughout winter. Thawing took place during May and June 2020 (“spring”) during which flow of MCW through BR2a and BR2b recommenced. The “summer” period included July and August 2020, and the “fall” September 2020. The results of the chemical parameters measured were averaged for each duplicate, i.e. BR1a and BR1b for BR1 and BR2a and BR2b for BR2. Genomic results were not averaged and are displayed for each individual BR.
The pH of the MCW feed varied from 7.2 to 8.5 throughout the experiment, with an average of 8.0 ± 0.4, with one outlier event measured at 9.3 (Figure S1). pH measured in the BRs were circum-neutral for the duration of their operating periods: BR1 pH averaged 7.4 ± 0.4, and BR2 pH averaged 6.9 ± 0.4. Conductivity measurements were used to estimate the concentration of total dissolved solids (Figure S2). The MCW used in the winter months had a higher conductivity (583 ± 47 mg/L average) than that used during the rest of the year (232 ± 92 mg/L average). After thawing (May 2020), effluent from BR2 had a higher conductivity than prior to freezing (October 2019) and until July 2020, BR2’s conductivity decreased and remained slightly below that of BR1. The major cation in the MCW was Ca2+, with concentrations ranging from 21.9 to 89.8 mg/L. The conductivity and Ca2+ concentration in the BR effluents tracked that of the MCW (Figure S2 and S3).
The temperature in BR1 varied from a high of 24.7°C on October 11, 2019 to a low of 6.2 ± 3.1°C averaged over the mid-October to mid-May period when the shed was heated (Figure S4). BR2’s highest temperature was 20.2°C also on October 11, 2019. BR1 and BR2 operated at an average temperature of 13.9 ± 2.3 and °C 13.4 ± 3.2°C, respectively from mid-May 2020 to the end of September 2020. ORP measurements (Table S1) taken every month ranged between − 105 ± 7 mV and − 348 ± 90 mV for BR1 and − 111 ± 3 mV and − 279 ± 19 mV for BR2. In BR1 the highest reducing conditions were reached during February 2020, while in BR2 it was on August 2020. Overall, the least reducing conditions were measured at the beginning of the experiment, during the start-up of the BRs.The average values were − 239 ± 90 mV and − 225 ± 62 mV for BR1 and BR2, respectively.
3.2 Sulfate reduction
Sulfate concentration in the MCW feed varied between 10 and 60 mg/L during fall 2019, spring and summer 2020. From mid-October 2019 to mid-January 2020, the concentration increased gradually in a semi-oscillatory fashion with a periodicity of about two weeks, corresponding to the bi-weekly refilling of the feed tank. The sulfate concentration in MCW remained stable, at 130 ± 6 mg/L, from mid-January to mid-April.
Sulfate concentration trends in the BR effluents followed those in the MCW, with higher concentrations during the winter months than during the spring, summer, and fall (Figure S5). Prior to winter, BR2 effluent sulfate concentrations (31 ± 7 mg/L) were higher than those in the BR1 effluents (15 ± 9 mg/L). However, after the spring thaw and throughout the summer and fall of 2020, both the indoor and outdoor bioreactor pairs achieved similar sulfate concentrations in their effluents (around 5 mg/L). Percentages of sulfate reduction efficiency (RE=[(In-Out)/In]x100)) were positive in all BRs over the entire experiment, except in five instances for BR1 and one for BR2 (Fig. 2). During winter, BR1 sulfate reduction efficiency was between 22.4 ± 4.6% and 65.5 ± 7.4%. In June, the average temperature increased to 14.1°C and sulfate reduction efficiency in BR1 increased to 81.4 ± 8.3%. From July 2020 until the end of the study, the sulfate reduction efficiency in BR1 remained high (> 95%). Prior to winter, 17 to 70% of sulfate was removed in BR2, while during the May thaw, sulfate removal was 50 ± 3%. After June 16 and until the end of the study, sulfate removal in BR2 was 88 ± 7%.
3.3 Carbon source consumption
The TOC measured in the MCW matched that predicted from the amount of molasses added (25 mg-C/L), with a few spike events detected mostly during the summer (Figure S6). Spike events in the MCW were due to external events (change in mining activities, weather, etc…) since the MCW used here, were sampled directly on the mine site.
From September to December 2019, the percentage of TOC removal was negative in all BRs (Figure S6 & S7), indicating release of carbon from a source inside the BRs. In the laboratory leaching experiments performed to explore this phenomenon, spruce wood chips leached 2.1 mg-C/g-wood-chips in column 1 and 1.5 mg-C/g-wood-chips in column 2 over the first week of their operation (Figure. S8). After five weeks, TOC leached from the spruce wood chips decreased to 0.6 mg-C/g-wood-chips in column 1 and 0.5 mg-C/ g-wood-chips in column 2 (Figure S8). The total amount of carbon leached from columns 1 and 2 were 6.5 and 5.0 mg-C/g-wood-chips (dry weight), respectively, over the five weeks of experiments. After December and until the end of the experiment, the average carbon consumption is positive (55 ± 16%) in BR1. When sampling of BR2 restarted during the spring and until the end of the experiment, TOC removal alternated negative and positive removal percentages (Figures S6 and S7).
3.5 Metal(loid)s
3.5.1. Iron leaching
Iron (Fe) concentrations in the MCW were less than 1.5 mg/L most of the time, except for eight peak events, with one reaching 35.2 mg/L (Figure S9). For the most part, concentrations of Fe were higher in the effluents than in the MCW, which indicates that Fe was being released from the BRs. One possible source of Fe was the inoculum. Wetlands are known to accumulate Fe (Doyle and Otte, 1997), and sediments used as inoculum were taken from a wetland close to the mine site.
3.5.2. Antimony
Sb concentrations in the MCW were between 0.48 and 5 µg/L during most of the study period (Figure S10). Two events with higher Sb concentrations (10.9 and 12.5 µg/L) were detected in April and August 2020, respectively. Sb concentration trends over time were similar in all effluents from BR1 and BR2, with lower concentrations than those of the MCW and varying between 0.41 ± 0.04 µg/L and 2.9 ± 1.1 µg/L over the study period.
Sb removal efficiency in BR1 was positive and greater than 20% most of the time, with the highest removal of 95.4% on August 6, 2020 (Fig. 3). Results from BR2 for Sb concentration were sparser due to sampling and analysis problems. Nevertheless, a maximum of 94.3% Sb was removed in BR2 on August 13, 2020. These results indicated that both BR1 and BR2 were similarly effective at Sb removal. Negative Sb removal efficiency was observed on few occasions, and was likely due to the time for BRs start up at the beginning of the experiment. Later on, negative Sb removal were observed when higher concentration of Sb was detected in the MCW, like in May 2020, and the negative Sb removal efficiency corresponds thus to the delay that the BRs needed to be able to remove this exceeding Sb.
3.5.3. Selenium
Over the year, Se in the MCW varied in concentration between 0.07 and 1.65 µg/L (Figure S11), with tendency to be slightly higher during spring and summer. Overall, Se concentrations in BR1 effluent were below those in the MCW and remained between 0.05 and 0.5 µg/L (Figure S11). Se removal in BR2 was similar to that observed in BR1, except after the spring thaw when the BR2 effluent Se concentrations (0.5 ± 0.1 µg/L) were greater than those in BR1 (0.12 ± 0.01 µg/L).
Se removal efficiency in BR1 was between 10 and 95% (Fig. 4). During its operating periods, BR2 achieved similar extents of Se removal, with the highest (93.8 ± 2.6%) obtained on the same date as that for BR1 (95.0 ± 0.6%). Just after the spring thaw, Se removal by BR2 increased to 53 ± 35%, but this was lower than BR1 over the same period (81 ± 2%). Two samplings in October and one at the beginning of November revealed higher Se concentrations in the effluents than in the MCW, indicating lower removal efficiency. The uncertainties on BR2 percentage of removal were higher during spring, i.e. during the thaw event, than for the rest of the study. It was likely due to the higher variability between BR2a and BR2b during the thawing period.
3.6 Bacterial population shifts
3.6.1 Overall microbial population composition and trends at a high taxonomic level: Phylum
For all BRs and consistently throughout the study period, the microbial population was constrained to 11 dominant phyla (Fig. 5), the most abundant being the Bacteroidota, Firmicutes, Desulfobacterota and Proteobacteria. Many members of the Desulfobacterota are known to perform sulfate reduction and this phylum was represented at the highest relative abundance in the indoor BRs (BR1a and b) during winter. They were at their lowest relative abundance in these same BRs the following summer. Overall, the prevalence of Desulfobacterota was lower post-spring 2020 in all BRs than previously. Desulfobacterota were more dominant in BR2 than BR1 post-spring 2020, but their percentage relative abundance varied to a greater degree in BR2 than in BR1.
Firmicutes were dominant throughout most BRs and seasons. These heterotrophic microorganisms are involved in the breakdown of complex carbon compounds through hydrolysis and fermentation. These might include compounds in the molasses fed to the BRs, dissolved organic carbon entering with the MCW feed, products from decomposing wood chips or carbon released from dead biomass (Sharmin et al. 2013; Liu et al. 2018). The Bacteroidota phylum had members known to be heterotrophic microorganisms, often known for scavenging partially degraded organic compounds (Mei et al. 2020a). Members of this group were present throughout the operation of the BRs.
It is known that sulfate-reducing bacteria, Firmicutes and Bacteroidota form syntrophic communities that cycle carbon. Sulfate-reducing bacteria, and other Desulfobacterota and Proteobacteria benefit from the degradation products produced by Firmicutes and Bacteriodota. It was observed in other studies of MCW treatment by BRs that the relative abundance of fermenters and sulfate-reducing bacteria vary cyclically over time out of phase with each other due to carbon cycling (Baldwin et al. 2016).
In a few instances, Proteobacteria were dominant, such as in BR2a during Fall 2019 and Fall 2020. These are functionally versatile microorganisms, some of which are involved in nitrogen cycling and/or oxidation-reduction of metals, which are important processes in BRs treating MCW (Johnson et al. 2019). The compositional differences between the indoor and outdoor BRs are not easily discernible at the phylum taxonomic level.
Variability in the bacterial population composition between the duplicate indoor and outdoor reactors was observed (Fig. 5). BR2a had a higher relative abundance of Desulfobacterota than BR2b. BR2b had compositionally more Firmicutes than Proteobacteria, whereas BR2a Proteobacteria were more prevalent. Bacterial population composition at the phylum level varied between the BR1 replicates during Fall 2019 and Winter; however, after the thaw, from Spring 2020 onwards, the microbial population composition in the two indoor BRs was more consistent.
Comparing changes in the population composition over the seasons (Fig. 5), the Desulfobacterota were much less dominant in the indoor BRs post-spring 2020 than in fall 2019/winter 2020. The overall population composition in the indoor BRs was consistent over time from spring 2020 onwards, whereas the population composition varied more in the outdoor bioreactors.
3.6.2 Microbial population differences and dominant ASVs
To study the bacterial population composition trends at a more granular level, a principal coordinate analysis plot of Bray-Curtis differences between sample ASV compositions was produced (Fig. 6). The only grouping of samples according to the indoor or outdoor BRs or seasons is the cluster of orange points on the left-hand side representing the indoor bioreactors from spring 2020 to fall 2020. During this period, the microbial population composition of these two BRs appeared to be stable. The same observation was not applicable for the other BRs.
To understand better the potential functional characteristics of the BR bacteria populations, the prevalence of dominant ASVs (> 1% of the total population) was studied (Fig. 7). Geobacter, an iron-reducing genus (Lovley et al. 2011), was the most dominant Desulfobacterota and was prevalent in the indoor BRs in the fall of 2019. Only one ASV was classified as a known sulfate-reducing bacteria genus, Desulfobulbus. It appeared to be consistently dominant in all BRs. The other Desulfobacterota were Citrifermentans and MSBL7. The former, also known as Geomonas, has one member that is a Fe-reducing species (Xu et al. 2019). MSBL7 is an uncharacterized environmental group. The most dominant Firmicutes ASVs were classified in an uncharacterized environmental group R-7 in the family Christensenellaceae, a known genus member of which are obligately anaerobic fermenters producing acetate and other organic acids from sugars (Morotomi et al. 2012). Sugar fermenting microorganisms were a major part of the bacterial communities in all BRs, as another dominant Firmicutes ASV was classified as Saccharofermentans (Chen et al. 2010). Of the Bacteroidota, the ASV classified as Bacteroidetes_vadinHA17 is thought to be a species adapted to the degradation of recalcitrant organic matter (Mei et al. 2020a). Finally, the other important phylum, Proteobacteria, contained three functionally versatile dominant ASVs associated with nutrient cycling and phototrophy.
Overall, the diversity of ASVs was constrained, with the number of ASVs in each sample varying from 100 to 350 (Table S2), which is within the range typical for MCW BRs but far less than the number of dominant ASVs in soils or vegetated wetlands, for example. The bubble plot of dominant ASVs in Fig. 7 is not very sparse, indicating that many ASVs were found in all BRs most of the time. The dominant ASVs were common among the BR populations, with differences being mostly due to compositional variation. The only exception was the Proteobacteria Stenotrophomonas ASV in BR2a_09/20. There are eight members of this genus, at the time of writing, some of which are known to be tolerant to high concentrations of heavy metals, and are important for N and S cycling (Ryan et al. 2009).
4.1. Tracking of bacterial activity and performance of the BRs
Throughout the experiment, the MCW fed into the BRs had circum-neutral to slightly basic pH, indicating neutral mine drainage conditions. Conductivity, Ca, metal(loid), TOC and sulfate concentrations in the MCW were lowest in fall 2019, increased in winter 2019–2020, and then decreased during the 2020 freshet. This trend in water chemistry is typical of seasonal variation: after the thawing period in late spring and throughout summer and early fall, the surface watershed in the Yukon is ice-free, and water flows freely, lowering conductivity and concentrations in MCW due to dilution. When temperatures are below freezing, as they are during winter, the volume of free water available on the surface is reduced, and thus, MCW is more concentrated in dissolved solids, explaining the observed increases in conductivity, Ca, metal(loid) and sulfate concentrations.
Calcium concentrations and conductivity measured in the BR effluents were at the same levels as those in MCW feed when considering an offset of two weeks, which is equal to the retention time of the MCW in the BRs. The pH also remained consistent between the MCW feed and BR effluents. Thus, it was concluded that (i) none of the Ca precipitated as sulfate minerals, such as gypsum or ettringite, in the BRs, and (ii) water chemistry parameters such as pH, Ca concentrations, conductivity were not strongly modified by the BRs.
Sulfate concentrations in the BR1 and BR2 effluents were lower than those in the MCW feed most of the time (87% and 91%, respectively), indicating that sulfate was being reduced in all BRs. Negative sulfate removal occurred only rarely (three occasions in BR1, and once in BR2) when the sulfate concentration in the MCW was higher than that measured in the effluents of BR1 and BR2. The negative sulfate removal values can be explained as time-lag or unsteady-state events that might have occurred when the influent sulfate concentration changed between two sampling events. Presence of SRB bacteria in the genus Desulfobulbus, which were observed in all BRs, supports the hypothesis that sulfate removal was occurring through biological reduction, with sulfate acting as an electron acceptor and molasses-derived carbon compounds as electron donors. In addition, optimal ORP values for SRB to reduce sulfate to sulfide that are in the range of -100 to -300 mV (Gibert et al. 2002), were measured in the BRs. In BR1, sulfate removal percentage appeared to correlate with temperature, with lower reduction observed during winter. This was expected since most SRB species are mesophiles, meaning that the optimal temperatures for their growth and activity are between 25 and 35°C, and similar temperature dependent observations were made in a previous BR study (Nielsen et al. 2018a). Sulfate reduction percentages were lowest in all BRs (< 50%) just after the thaw period. Nevertheless, sulfate reduction in all BRs increased to over 80% within a month after thawing, and was maintained until the end of the study. These results indicate that SRB re-established their bacterial activity quickly after thaw with only a small delay between the increase in temperature and the increase in bacterial activity. Since the SRB related ASVs present before and after winter were taxonomically similar, the SRB active from spring 2020 onwards could have been those that survived freeze/thaw inside BR2. It is unknown what microorganisms were present in the MCW or how consistent this was before and after winter. If different microbial species entered the BRs with the spring freshet MCW, they did not change the taxonomy of the SRB population.
Another tracker of bacterial activity in the BRs was carbon removal, which was expected due to consumption by active bacteria. The presence of bacteria involved in the decomposition of organic C compounds, and other bacteria (SRB and iron-reducing bacteria) that are heterotrophic, suggests that carbon source consumption must be taking place. However, for the first four months of the experiment, carbon was released rather than removed in BR1. Leaching experiments using the same spruce wood chips as were used in the BRs revealed that up to 6.5 mg-C/g woodchips (dry weight) were leached over five weeks. Based on these results, it is likely that soluble carbon compounds were being released from the spruce wood chips in the BRs, at least during the initial period after start-up. During this period, the amount of carbon released was greater than the carbon consumed by bacterial activity. Presence of microbial taxa associated with degradation of recalcitrant organic material, such as hard to breakdown lignocellulosic compounds (i.e. Bacteroidetes_vadinHA17 (Mei et al. 2020b)), provides evidence for the biological decomposition of the wood chips in the BRs. After four months, the percentage of carbon removal in BR1 became positive, but the values remained negative for BR2 until the end of the experiment. A likely explanation for this was that biodegradation of wood chips occurred to a greater extent in the indoor BRs since they operated over the entire experimental period, whereas freezing of the outdoor BRs prevented any bacterial activity for enzymatic hydrolysis of cellulosic material in the woodchips during winter. Heterotrophic bacteria such as the Firmicutes Saccharofermentas were dominant throughout all BRs and seasons but were present at higher relative abundances in BR2 than BR1 in the summer and fall of 2020. Members of this genus were likely growing on the sugars still being produced in BR2 from enzymatic hydrolysis of the woodchips. Decomposition of wood chips inside the BRs might be beneficial since this would increase the supply of carbon for bacterial growth. However, this supply will diminish over time as all the degradable C is depleted, which seemed to be the case for BR1. Thus, extraneous carbon addition in the form of cost-effective natural sources such as molasses is needed for long term reliable performance of BRs.
4.2. Impact of bacterial activity on metalloids removal
While Sb concentrations in the MCW did not vary seasonally, Se levels depended on the season, and were highest in summer (Figure S11). No strong correlations (R ≤ 0.9) were detected between Sb or Se and major elements such as sulfate, carbon, Fe, Ca or other metals (Pb, Cu, Cd, Pb) in the MCW. After one month of operation, in October 2019, the BRs started to remove Sb and Se, coinciding with signs of sulfate reduction. Sb likely entered the BRs as Sb(V) since the MCW used in this study was collected at the surface (Johnston et al. 2020). Processes that might be involved in Sb removal in the reducing environment of the BRs include microbial Sb(V) reduction to Sb(III), microbial methylation, and or abiotic Sb(V) reduction by S(-II) and or Fe(-II) (Zhang et al. 2022). Microbial Sb(V) reduction was also likely occurring in the BRs. This can take place through respiratory and non-respiratory metabolic pathways. For instance, Sb(V) can be enzymatically reduced by SRB for energy (Wang et al. 2013), with the reduced Sb immobilized via antimony sulfide formation (Liu et al. 2022). Other types of bacteria have been reported for mediating immobilization and removal of Sb, under both aerobic and anaerobic conditions, including those that oxidize the more toxic Sb(III) to the less toxic Sb(V) (Deng et al. 2021). None of the Sb resistant genera reported in the Deng et al. 2021 study were among those identified in the BRs. Biological Sb methylation can occur under anoxic conditions as a detoxification mechanism and has been observed in various environments, but the molecular mechanism remains poorly understood, and only a few bacterial species with this capability have been identified so far (Liu et al. 2022). None of the latter genera reported in Liu et al. 2022 were identified the BRs. Another path to Sb(V) removal is through SRB species(Wang et al., 2013). Of all the possible Sb biotic removal mechanisms, we hypothesize that biotic removal mediated by SRB was very likely occurring in the BRs due to the prevalence of Desulfobulbus species. The presence of Fe-reducing bacteria in the BRs, the reducing conditions in the BRs(Jia et al. 2022) and the higher concentrations of Fe compared to Sb (mg/L vs. µg/L) suggested that Sb-Fe oxide precipitation could occurred in the BRs, despite the continuous release of Fe in the effluents. Abiotic precipitation of antimony sulfide was another possible Sb removal mechanism, given the presence of sulfide (HS−, S2−) coming from sulfate reduction.
Overall, both the indoor and outdoor BRs were successful at removing Se from the MCW. The removal percentage was between 10 and 95% for BR1 and 44 and 95% for BR2. Prior to freezing, BR2 removed similar amounts of Se as BR1 (around 60%). Post winter, during the thaw period (May 2020), BR2 recommenced removing Se, however at lower percentages (53 ± 35%) than in BR1 (81 ± 2%) during the same period. By summer 2020, both BR1 and BR2 removed similar amounts of Se. These results indicate that while BR2’s Se removal capacity was impacted by freezing, the BR recovered quickly achieving the same removal capacity as BR1 after only a few weeks.
Se speciation was not performed, and so it is not known which chemical forms were present in the MCW or in the BR’s effluents. As the MCW was sourced at the surface where conditions were oxidizing, then Se was present likely as selenate (SeO42−). Many bacteria can reduce selenate as an electron acceptor for energy and growth. These include those with selenate specific reductases, such as Thauera selenatis (Macy et al. 1993), or denitrifying bacteria with nitrate reductases that are versatile (e.g. Paracoccus sp.) (Sabaty et al. 2001). Selenate or selenite (SeO32−) reduction can occur aerobically or anaerobically (Avendaño et al. 2016). None of the species that have been characterized at the time of publishing as Se reducing bacteria were among the dominant AVS genera identified in the BR metagenomes. However, Se removal has been observed in SRB bioreactors in the past mediated through SRB activity (Hockin and Gadd 2006). Due to the presence of SRB in the BRs, the latter was a probable mechanism. Selenate is the most soluble and mobile form of Se. When selenate is reduced to selenite, which is another soluble oxyanion of Se with valence Se(IV), then this chemical form has a higher affinity for adsorption than selenate. If selenate reduction to selenite was occurring in the BRs, then Se removal might have occurred through selenite binding to solids such as Fe oxy(hyrox)ides (Balistrieri and Chao 1990).
Relative abundances of the microbial ASVs in BR1 were more stable across time points post the Spring 2020 thaw than that in BR2. The greater relative abundance variability observed in BR2 versus BR1 during this period is attributed to the effects of freeze-thaw in BR2. Nevertheless, many of the same dominant genera were present both pre- and post-freeze-thaw, suggesting that biological removal mechanisms for Se and Sb were consistent across seasons. Since the microbial population included phylotypes with assumed phenotypes with metabolic potential for C, sulfate, Se and Sb removal and Fe release, the successful performance of all BRs was not limited by a functionally essential microbial community. Presence of this microbial community might have contributed to the relatively quick re-establishment of BR2’s performance post thaw. Indeed, other studies also have made the important observation that distribution of key metabolic functions across different genera coexisting in the BRs favours process stability (Ayala-Muñoz et al. 2021). Overall, this study demonstrated that freezing of BRs did not substantially affect their ability to remove sulfate, Se and Sb from MCW.
The present study provides evidence that BRs inoculated with locally sourced bacterial populations, maintain their capability to remove sulfate, as well as Sb and Se metalloids despite being subject to a seasonal freeze-thaw cycle. BRs that were located inside a heated shed, maintained at 5°C at least, consistently removed Se and Sb over the whole year. BRs that were located outdoors removed similar amounts of Se and Sb as the BRs inside the shed for the periods prior to freezing. After the spring thaw, removal percentages of Se and Sb in the outdoor BRs were slightly lower than those achieved in BR1. However, after thawing and during subsequent summer and fall, Se and Sb removal was similar in both indoor and outdoor bioreactors. Microbial population composition analysis indicated that taxonomic groups aligned with expected biological metalloids removal mechanisms were present. These included bacteria involved in iron reduction, sulfate reduction and decomposition of organic material. The presence of Desulfobulbus genera and based on evidence in previous studies, the most likely postulated mechanism for Se and Sb removal was mediation by SRB either directly through dissimilatory reduction or indirectly through sulfide precipitation. The freeze-thaw events over winter and spring did not disrupt the microbial population composition with similar dominant bacteria being present before and after. Using inoculum sourced directly on site might explain these interesting and positive observations: The bacterial consortium growing inside the BRs contains bacteria naturally present in the (sub)arctic region and, thus, adapted to freezing events occurring during winter. One pitfall of using anaerobic BRs for MCW treatment is the release of Fe in BRs’ effluents. Fe, such as other chemical elements, is a regulated parameter in the mining industry, and its release in the environment following water treatment needs to be considered when designing future semi-passive water treatment systems. Further studies focusing on several freeze-thaw cycles, various inoculums, and different contaminants would allow us to extend and validate the promising results obtained in this study.
ASV: Amplicon Sequence Variant
SRB: Sulfate-Reducing Bacteria
BR: Bioreactor
MCW: Mine-Contact Water
Se: Selenium
Sb: Antimony
°C: degree Celsius
HRT: hydraulic retention time
Fe: Iron
Ca: Calcium
Cd: Cadmium
Cu: Copper
Acknowledgements (including funding sources)
The Natural Sciences and Engineering Research Council of Canada (NSERC) is acknowledged for supporting this work through the Industrial Research Chair in Northern Mine Remediation at Yukon University.
Victoria Gold Staff are acknowledged for supporting the fieldwork. This pilot scale experiment would not have been possible without daily monitoring and maintenance that was performed by the Victoria Gold environmental team.
- Allison SD, Martiny JBH (2008) Resistance, resilience, and redundancy in microbial communities. Proceedings of the National Academy of Sciences 105:11512–11519
- Avendaño R, Chaves N, Fuentes P, et al (2016) Production of selenium nanoparticles in Pseudomonas putida KT2440. Sci Rep 6:37155
- Ayala‐Muñoz D, Simister RL, Crowe SA, et al (2021) Functional redundancy imparts process stability to acidic Fe (II)‐oxidizing microbial reactors. Environ Microbiol 23:3682–3694
- Baldwin SA, Mattes A, Rezadehbashi M, Taylor J (2016) Seasonal microbial population shifts in a bioremediation system treating metal and sulfate-rich seepage. Minerals 6:36
- Balistrieri LS, Chao TT (1990) Adsorption of selenium by amorphous iron oxyhydroxide and manganese dioxide. Geochim Cosmochim Acta 54:739–751
- Bolyen E, Rideout JR, Dillon MR, et al (2019) Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nat Biotechnol 37:852–857
- Callahan BJ, McMurdie PJ, Rosen MJ, et al (2016) DADA2: High-resolution sample inference from Illumina amplicon data. Nat Methods 13:581–583
- Chen S, Niu L, Zhang Y (2010) Saccharofermentans acetigenes gen. nov., sp. nov., an anaerobic bacterium isolated from sludge treating brewery wastewater. Int J Syst Evol Microbiol 60:2735–2738
- Coulomb R, Dietz S, Godunova M, Nielsen TB (2015) Critical minerals today and in 2030: an analysis for OECD countries
- Deng R, Chen Y, Deng X, et al (2021) A critical review of resistance and oxidation mechanisms of Sb-oxidizing bacteria for the bioremediation of Sb (III) pollution. Front Microbiol 12:738596
- Doyle MO, Otte ML (1997) Organism-induced accumulation of iron, zinc and arsenic in wetland soils. Environmental Pollution 96:1–11
- Gibert O, de Pablo J, Cortina JL, Ayora C (2002) Treatment of acid mine drainage by sulphate-reducing bacteria using permeable reactive barriers: a review from laboratory to full-scale experiments. Rev Environ Sci Biotechnol 1:327–333
- Habe H, Sato Y, Aoyagi T, et al (2020) Design, application, and microbiome of sulfate-reducing bioreactors for treatment of mining-influenced water. Appl Microbiol Biotechnol 104:6893–6903
- Haley S, Klick M, Szymoniak N, Crow A (2011) Observing trends and assessing data for Arctic mining. Polar Geography 34:37–61
- Hockin S, Gadd GM (2006) Removal of selenate from sulfate‐containing media by sulfate‐reducing bacterial biofilms. Environ Microbiol 8:816–826
- Jia X, Ma L, Liu J, et al (2022) Reduction of antimony mobility from Sb-rich smelting slag by Shewanella oneidensis: Integrated biosorption and precipitation. J Hazard Mater 426:127385
- Johnson DB, Hallberg KB (2005) Acid mine drainage remediation options: a review. Science of the total environment 338:3–14
- Johnson H, Cho H, Choudhary M (2019) Bacterial heavy metal resistance genes and bioremediation potential. Comput Mol Biosci 9:1
- Johnston SG, Bennett WW, Doriean N, et al (2020) Antimony and arsenic speciation, redox-cycling and contrasting mobility in a mining-impacted river system. Science of the Total Environment 710:136354
- Leppänen JJ, Weckström J, Korhola A (2017) Multiple mining impacts induce widespread changes in ecosystem dynamics in a boreal lake. Sci Rep 7:10581
- Liu H, Sun W, Häggblom MM (2022) Microbial transformations of antimony. In: Microbial metabolism of metals and metalloids. Springer, pp 223–254
- Liu L, Wang S, Guo X, et al (2018) Succession and diversity of microorganisms and their association with physicochemical properties during green waste thermophilic composting. Waste Management 73:101–112
- Lovley DR, Ueki T, Zhang T, et al (2011) Geobacter: the microbe electric’s physiology, ecology, and practical applications. Adv Microb Physiol 59:1–100
- Macy JM, Rech S, Auling G, et al (1993) Thauera selenatis gen. nov., sp. nov., a member of the beta subclass of Proteobacteria with a novel type of anaerobic respiration. Int J Syst Evol Microbiol 43:135–142
- Martin AJ, Jones R, Buckwalter-Davis M (2009) Passive and semi-passive treatment alternatives for the bioremediation of selenium from mine waters
- McNulty BA, Jowitt SM (2021) Barriers to and uncertainties in understanding and quantifying global critical mineral and element supply. iScience 24:102809
- Mei R, Nobu MK, Narihiro T, Liu W-T (2020a) Metagenomic and metatranscriptomic analyses revealed uncultured bacteroidales populations as the dominant proteolytic amino acid degraders in anaerobic digesters. Front Microbiol 11:593006
- Mei R, Nobu MK, Narihiro T, Liu W-T (2020b) Metagenomic and metatranscriptomic analyses revealed uncultured bacteroidales populations as the dominant proteolytic amino acid degraders in anaerobic digesters. Front Microbiol 11:593006
- Morotomi M, Nagai F, Watanabe Y (2012) Description of Christensenella minuta gen. nov., sp. nov., isolated from human faeces, which forms a distinct branch in the order Clostridiales, and proposal of Christensenellaceae fam. nov. Int J Syst Evol Microbiol 62:144–149
- Muyzer G, Stams AJM (2008) The ecology and biotechnology of sulphate-reducing bacteria. Nat Rev Microbiol 6:441–454
- Neculita C, Zagury GJ, Bussière B (2007) Passive treatment of acid mine drainage in bioreactors using sulfate‐reducing bacteria: Critical review and research needs. J Environ Qual 36:1–16
- Ness I, Janin A, Stewart K (2014) Passive treatment of mine impacted water in cold climates: a review. Yukon Research Centre, Yukon College 56
- Nielsen G, Hatam I, Abuan KA, et al (2018a) Semi-passive in-situ pilot scale bioreactor successfully removed sulfate and metals from mine impacted water under subarctic climatic conditions. Water Res 140:268–279
- Nielsen G, Janin A, Coudert L, et al (2018b) Performance of sulfate-reducing passive bioreactors for the removal of Cd and Zn from mine drainage in a cold climate. Mine Water Environ 37:42–55
- Parada AE, Needham DM, Fuhrman JA (2016) Every base matters: assessing small subunit rRNA primers for marine microbiomes with mock communities, time series and global field samples. Environ Microbiol 18:1403–1414
- Price PB, Sowers T (2004) Temperature dependence of metabolic rates for microbial growth, maintenance, and survival. Proceedings of the National Academy of Sciences 101:4631–4636
- Ryan RP, Monchy S, Cardinale M, et al (2009) The versatility and adaptation of bacteria from the genus Stenotrophomonas. Nat Rev Microbiol 7:514–525
- Sabaty M, Avazeri C, Pignol D, Vermeglio A (2001) Characterization of the reduction of selenate and tellurite by nitrate reductases. Appl Environ Microbiol 67:5122–5126
- Sengupta M (2021) Environmental impacts of mining: Monitoring, restoration and control. CRC Press
- Sharmin F, Wakelin S, Huygens F, Hargreaves M (2013) Firmicutes dominate the bacterial taxa within sugar-cane processing plants. Sci Rep 3:3107
- Tiainen H, Sairinen R, Sidorenko O (2015) Governance of Sustainable Mining in Arctic Countries: Finland, Sweden, Greenland & Russia. Arctic yearbook 2015:132
- Walters W, Hyde ER, Berg-Lyons D, et al (2016) Improved bacterial 16S rRNA gene (V4 and V4-5) and fungal internal transcribed spacer marker gene primers for microbial community surveys. mSystems 1:e00009-15
- Wang H, Chen F, Mu S, et al (2013) Removal of antimony (Sb(V)) from Sb mine drainage: Biological sulfate reduction and sulfide oxidation–precipitation. Bioresour Technol 146:799–802. https://doi.org/https://doi.org/10.1016/j.biortech.2013.08.002
- Warnow T (2015) Encyclopedia of Metagenomics. Springer US, Boston, MA
- Wright IA, McCarthy B, Belmer N, Price P (2015) Subsidence from an underground coal mine and mine wastewater discharge causing water pollution and degradation of aquatic ecosystems. Water Air Soil Pollut 226:1–14
- Xu Z, Masuda Y, Itoh H, et al (2019) Geomonas oryzae gen. nov., sp. nov., Geomonas edaphica sp. nov., Geomonas ferrireducens sp. nov., Geomonas terrae sp. nov., Four Ferric-Reducing Bacteria Isolated From Paddy Soil, and Reclassification of Three Species of the Genus Geobacter as Members of the Genus Geomonas gen. nov. Front Microbiol 10:2201
- Younger PL, Banwart SA, Hedin RS, et al (2002) Mine water hydrology. Springer
- Zhang Y, O’Loughlin EJ, Kwon MJ (2022) Antimony redox processes in the environment: A critical review of associated oxidants and reductants. J Hazard Mater 128607
- Zhao Y-G, Wang A-J, Ren N-Q (2010) Effect of carbon sources on sulfidogenic bacterial communities during the starting-up of acidogenic sulfate-reducing bioreactors. Bioresour Technol 101:2952–2959